Riskiest Assumption

The belief that kills the company if wrong, which you should test first.

Also known as: riskiest assumption, leap of faith assumption, critical assumption

The riskiest assumption is the belief your company most depends on and knows least about — wrong means death, right means little (other risks remain). Everything else can wait: test this one first, with the cheapest experiment that yields real evidence.

list assumptions → rank by (dies-if-wrong × uncertainty) → test #1 cheapest way possible
e.g. "restaurants will pay" beats "the UI should be blue" by demolition order

Founders reliably test comfortingly testable beliefs (logo, stack, name) while the killer belief (will anyone pay? can we acquire cheaply? does the regulation allow it?) sits untouched for months. Name yours in one sentence; if you cannot, that paralysis is itself the finding.

The classic mistakes:

Testing in comfort order. Easy tests first feels productive and teaches nothing load-bearing. Discomfort is the compass — the test you avoid is the test to run.
One big test for a compound belief. “People want this and will pay $X and we can reach them for $Y” is three assumptions wearing a trench coat. Split and test separately, cheapest first.
Mistaking opinions for evidence. Ten “great idea!” responses test politeness. Evidence is behavior: money, signups with effort, data shared, pilots signed (see pre-selling).
Retiring a risk prematurely. One successful test downgrades a risk; it rarely eliminates it. Keep the ranked list alive and re-rank as evidence arrives.

The weekly question: “What is most likely to kill us that we still believe without evidence?” Work on that, and only that, until it flips.